Samarium Monosulfide (SmS): Reviewing Properties and Applications.

materials Review

Samarium Monosulfide (SmS): Reviewing Properties and Applications Andreas Sousanis, Philippe F. Smet

ID

and Dirk Poelman *

ID

Lumilab, Department of Solid State Sciences, Ghent University, Krijgslaan 281/S1, 9000 Gent, Belgium; [email protected] (A.S.); [email protected] (P.F.S.) * Correspondence: [email protected]; Tel.: +32-9-264-4367 Received: 5 May 2017; Accepted: 10 August 2017; Published: 16 August 2017

Abstract: In this review, we give an overview of the properties and applications of samarium monosulfide, SmS, which has gained considerable interest as a switchable material. It shows a pressure-induced phase transition from the semiconducting to the metallic state by polishing, and it switches back to the semiconducting state by heating. The material also shows a magnetic transition, from the paramagnetic state to an antiferromagnetically ordered state. The switching behavior between the semiconducting and metallic states could be exploited in several applications, such as high density optical storage and memory materials, thermovoltaic devices, infrared sensors and more. We discuss the electronic, optical and magnetic properties of SmS, its switching behavior, as well as the thin film deposition techniques which have been used, such as e-beam evaporation and sputtering. Moreover, applications and possible ideas for future work on this material are presented. Our scope is to present the properties of SmS, which were mainly measured in bulk crystals, while at the same time we describe the possible deposition methods that will push the study of SmS to nanoscale dimensions, opening an intriguing range of applications for low-dimensional, pressure-induced semiconductor–metal transition compounds. Keywords: SmS; switchable material; thin films; magnetic properties; optical properties; deposition techniques

Switchable materials are a specific category of compounds presenting changes between two distinct states, with a simultaneous change of material properties. Several kinds of switchable materials showing a semiconductor–metal transition are known. In phase change materials such as AgInSbTe and Ge2 Sb2 Te5 , there is a reversible transition between an amorphous and a crystalline phase, by using either optical or electrical pulses, changing both the structure of the initial material and the corresponding optical and electrical properties. These materials can be used in non-volatile memory devices, exploiting characteristics from flash memories and dynamic random access memories, respectively [1–3]. Vanadium dioxide (VO2 ) is another example of a switchable material, which could be exploited in several applications, such as thermochromic windows [4–7]. In this case, the transition is related to a thermally induced change between the rutile and tetragonal phase, which is accompanied by changes in its properties. Samarium monosulfide (SmS) is known for its pressure-induced semiconductor–metal transition, leading to a plethora of exciting physical properties. In this review, we will discuss some of the most important results about SmS, showing possibilities to use it as a material for a new kind of non-volatile memory, which could exploit the pressure-induced transition of SmS. This kind of application requires new research on material properties at low dimensions. This review first describes bulk material properties but also attempts to stimulate further exploitation and study of SmS thin and ultra-thin films.

Materials 2017, 10, 953; doi:10.3390/ma10080953

www.mdpi.com/journal/materials

Materials 2017, 10, 953

2 of 21

1. Introduction SmS is an intriguing material, which shows a pressure-induced transition from the semiconducting to the metallic state [8–10]. The transition is also called the black–golden transition, referring to the typical color of the SmS in the semiconducting and metallic state, respectively. Generally, the transition appears at about 6.5 kbar, at room temperature [8,11–13]. The importance and uniqueness of this material is based on its switching characteristics, as other switchable materials require proper heating/cooling cycles, such as VO2 , or only show a metal–insulator transition at low temperatures, due to the splitting of the f-band in strongly correlated systems. As we will discuss further, the moderate pressure required for switching, the possibility to modify the material to make the transition reversible, and finally the fact that the transition takes place at room temperature, all make SmS a highly interesting material from the applications point of view. From the first investigations by Suryanarayanan et al. [14] to more recent work of Rogers et al. [15] the optical spectra of SmS have been studied in order to gain a deeper understanding of the absorption peaks and the nature of the transition. SmS possesses a NaCl (fcc/rock salt) structure [13,15], which remains unaltered after transition, as seen in Figure 1a. The isostructural transition contrasts with other phase change materials, where changes in resistivity and optical properties are caused by a change in crystallographic phase. The process that takes place in SmS is a volume collapse, with a decrease in unit cell volume by approximately 13.5% [16]. The semiconducting phase can be restored by thermal annealing. By alloying with europium (Eu) or gadolinium (Gd), the pressure at which the phase transition occurs can be changed [17,18]. The Mott criterion gives the concentration of electrons required for the electronic transition to occur, independent of the method to induce the transition, such as light, heat or pressure [10,19,20]. It has been found that, in the case of a laser-induced non-thermal semiconductor–metal transition, the minimum density of electrons upon the transition should approach the electron density of the metallic state [20,21]. In Figure 1a (upper left graph), we see the main X-ray diffraction (XRD) peak of a semiconducting SmS thin film and its shift to larger angles in the pressure induced metallic phase [15]. This shift indicates a decrease in lattice constant, which can be used as a probe for the transition to the metallic state. Moreover, theoretical models such as the theory of excitonic instability can predict and explain this transition, as presented below. The semiconducting–metallic transition of SmS is driven by the variable oxidation state of Sm. Samarium has two possible ground state configurations in SmS, a nonmagnetic Sm2+ with 4f6 (7 F0 ) configuration and a magnetic Sm3+ with 4f5 (6 H5/2 ) configuration [22]. It has to be noted that an intermediate valence state is also possible. Other mixed valence Sm compounds are SmSe, SmTe and SmB6 [12,23]. We can observe the changes in valence state of SmS in X-ray absorption near edge structure (XANES) measurements, as shown in Figure 1a (lower left graph). The Sm valence state changes from a predominantly 2+ state (black, semiconducting state) to a mainly 3+ state (golden, metallic state), with a remaining shoulder at the position of Sm2+ , related to the existence of a mixed valence state or to the presence of non-switched parts in the thin film. In order to quantify these results, least-squares fitting was used, showing an intermediate valence, even above 2 GPa, where the magnetic order appears. More details on this behavior are given below. In addition, computations showed that Sm does not acquire its fully trivalent state even at a pressure higher than 6 GPa. The lattice constant decreases as the valence state increases, due to the large difference in ionic radius for Sm2+ and Sm3+ upon the transition to the metallic state. In Figure 1b, we can see the behavior of the SmS single crystal resistivity as a function of hydrostatic pressure. In general, the resistivity drop occurs at close to 6.5 kbar (at room temperature), but there are small differences depending on the experimental measurement conditions [24]. The resistivity as a function of pressure can be expressed as: ρ P = ρ0 e−(α) P/kT , where α = d(∆Eg )/dP is the pressure dependent gap between the f-d bands. In the case of a continuous transition of SmS, the ambient pressure energy gap ∆Eg can be found from the ρ a /ρh , where ρ shows the resistivity at ambient pressure, while ρh represents the resistivity at high pressure, since ρh = ρ a e−∆Eg /kT . The energy gap between the f and d band goes to zero, before the transition [25].

Materials 2017, 10, 953 Materials 2017, 10, 953

3 of 21 3 of 21

Figure1. 1. (a) (a) Basic Basic properties properties of of SmS. SmS. The The XRD XRD pattern pattern (upper (upper left) left) demonstrates demonstrates the the shift shift of of the the (200) (200) Figure diffractionpeak, peak,when whenS-SmS S-SmS(semiconducting, (semiconducting,black blackcurve) curve)switches switchesto toM-SmS M-SmS(metallic). (metallic).The Thelower lower diffraction left graph graphshows showsthe thecorresponding corresponding valence valence change, change, as asevident evidentfrom fromaaXANES XANESmeasurement. measurement. In In the the left S-SmSor orblack black state state the the material material shows shows aa NaCl structure structure with a black-blue color and a high resistivity. S-SmS resistivity. TheM-SmS M-SmSor orgolden goldenstate statehas hasthe thesame samestructure structurewith withaavolume volumecollapse. collapse.The Thecolor colorchanges changesto togold gold The andthe theresistivity resistivitydrops dropsup upto tofour fourorders ordersof ofmagnitude. magnitude. For For the the unit unit cell cell of of SmS: SmS: purple purple spheres, spheres, Sm; Sm; and yellowspheres, spheres,S.S. Resistivity drop of SmS single crystal. Four different pressure transmitting yellow (b)(b) Resistivity drop of SmS single crystal. Four different pressure transmitting media media are indicated in the (Reprinted graph. (Reprinted with permission from [24], respectively.) are indicated in the graph. with permission from [26], [27][26], and[27] [24],and respectively.)

Upto to this this point, point, we we briefly briefly mentioned mentioned the the basic basic properties properties and and switching switching characteristics characteristics of of SmS, SmS, Up aswell wellas as ways ways to to monitor monitor these these properties. properties. In In the the next next sections, sections, we we will will describe describe the the pressure pressure related related as electronic, optical and magnetic properties of SmS in a more detailed way, as they were typically electronic, optical and magnetic properties of SmS in a more detailed way, as they were typically studied in in bulk bulk crystals. crystals. While While interesting interesting in in itself, itself, the the utilization utilization of of these these properties properties in in real-world real-world studied applications requires high quality SmS thin films. As the deposition of stoichiometric SmS filmswith with applications requires high quality SmS thin films. As the deposition of stoichiometric SmS films the correct phase and Sm valence state is far from obvious, part of this review will be devoted to the correct phase and Sm valence state is far from obvious, part of this review will be devoted to describingthe thedifferent differentmethods methodsto tosynthesize synthesizeSmS SmSthin thinfilms. films. describing 2. Properties Properties of of SmS SmS 2. 2.1. Energy Energy Level Level Diagram Diagram 2.1. Electronicstructure structureand andpressure pressure related properties of SmS were reviewed in 1993 by Wachter Electronic related properties of SmS were reviewed in 1993 by Wachter [28], [28], essentially onSmS. bulkInitially, SmS. Initially, theproperties, basic properties, such as piezoresistivity andchanges optical essentially on bulk the basic such as piezoresistivity and optical changes beenwhile studied, while of the Kondo-like behavior and laser-induced transition have beenhave studied, aspects of aspects the Kondo-like behavior and laser-induced transition properties properties have been investigated over the The last years. The maindifference electronic between difference between SmS have been investigated over the last years. main electronic SmS and other 6 2+ 5 6 2+ and other chalcogenides is the the gap between the 4f ground theand Smtheion chalcogenides is the fact that thefact gapthat between the 4f ground state of thestate Sm ofion 4f and 5d(tthe 2g ) 4f55d(t2g) degenerate state is in smaller in comparison withsimilar other similar materials [12,15,29]. In Figure 2a, degenerate state is smaller comparison with other materials [12,15,29]. In Figure 2a, the the band structure of three Sm chalcogenides is shown, as calculated by Batlogg et al. [30]. Recently, band structure of three Sm chalcogenides is shown, as calculated by Batlogg et al. [30]. Recently, it it has been reported that the energy betweenthe thep-character p-characterofofthe thevalence valenceband bandand andthe the4f 4fstates states of of has been reported that the energy between the samarium samarium ions ions is is constant constant [31]. [31]. By By pressurizing pressurizing SmS, SmS, the the 5d 5d degenerate degenerate state state moves moves toward toward the the the 2+ ion, because the increase in 4f66 ground ground state state of of the the Sm Sm2+ an increase increase of of the the energy energy 4f pressure leads to an splittingofofthe thedoubly doublydegenerate degenerate ) and triply degenerate states [32,33]. In Figure 2b,can we splitting (eg(e) gand triply degenerate (t2g(t ) 2g 5d) 5d states [32,33]. In Figure 2b, we can observe that the band structure separates into three parts, as shown in the density of states (DOS). observe that the band structure separates into three parts, as shown in the density of states (DOS). The The 5d SmS 5d conduction band eV the above thelevel, Fermithe level, thelocalized rather localized Sm 4f are SmS conduction band is 0.25iseV0.25 above Fermi rather Sm 4f levels arelevels situated situated 0between 0 and eVhighly and the highly dispersive S 3p valence bandbelow is located below eV between and −1 eV and−1 the dispersive S 3p valence band is located −1.6 eV [34].−1.6 There [34]. There is an indirect gap between the Γ point of the valence band and the bottom of the is an indirect gap between the Γ point of the valence band and the bottom of the conduction band at conduction band at point, equal tothe 0.25 eV. gap Moreover, the X point, equal to the 0.25XeV. Moreover, direct is equalthe to direct 0.5 eV.gap is equal to 0.5 eV.


4 of 21 4 of 21

Figure2.2.(a) (a)Band Bandstructures structures semiconducting monochalcogenides SmSe and SmTe, Figure forfor semiconducting monochalcogenides SmS,SmS, SmSe and SmTe, wherewhere D(E) is the density of For states. Forcases, thesethe cases, the gap between and unoccupied states is isD(E) the density of states. these gap between occupiedoccupied and unoccupied states is different occupied;from (b)computational Band structure from different toaccording to Batlogg, with the 4f5 state according Batlogg, with the 4f5 state fully occupied; (b) fully Band structure work, in the semiconducting state following the LSDA+U energy of Sm 4f) methodenergy with computational work, in the semiconducting state(Coulomb following repulsion the LSDA+U (Coulomb repulsion the contribution of with spin-orbit coupling. (Reprinted with permission from [30] andpermission [34], respectively.) of Sm 4f) method the contribution of spin-orbit coupling. (Reprinted with from [30] and [34], respectively.)

2.2. Electrical and Optical Properties 2.2. Electrical and Optical Properties Here, the behavior of the dielectric function in both semiconducting and metallic states is Here, There the behavior the dielectric function in both semiconducting SmS and between metallic 0.8 states presented. are fiveofprominent optical transitions in semiconducting andis presented. There are five prominent optical transitions in semiconducting SmS between 0.8 and 4 eV, 4 eV, as observed by UV/VIS spectroscopy [15,35] and angle-resolved photo-emission spectroscopy − 1 6 7 5 6 as observed by UV/VIS spectroscopy [15,35] and angle-resolved photo-emission spectroscopy (ARPES) [32]. The first transition with a maximum at 0.8 eV (6650 cm ) is related to 4f ( F0 )-4f ( H)5d 1 (6650 cm−1) is related 6 7to 4f6(75F60)-4f5(6H)5d [32]. The first transition with a maximum at 0.8−eV (t(ARPES) 2g ); the second transition, located at 1.6 eV (13,215 cm ), originates from 4f ( F0 )-4f ( F)5d (t2g ); 6F)5d (t2g); and 1 ) corresponds (t2g);the thethird second transition, located at 1.6cm eV−(13,215 cm−1), originates 4f65((76FH)5d 0)-4f5((e and transition at 3.1 eV (25,000 to the 4f6from (7 F0 )-4f g ) transition. −1 6 7 5 6 6 7 5 6 5 6 the other third two transition at 3.1take eV (25,000 cm 4f ) corresponds the 4f ( F 0)-4f ( H)5d (e g)) and transition. The transitions place from ( F0 ) to 4f ( to P)5d(t ) and 4f ( F)5d(e appearThe at g 2g other 3.6 twoeV transitions take place from 4f6Figure (7F0) to 3b). 4f5(6P)5d(t 2g) and 4f5(6F)5d(e g) and about 3.6 about and 4.0 eV, respectively (see The basic knowledge about theappear opticalatproperties eVSmS andcomes 4.0 eV, respectively Figure 3b). The knowledge about the optical of SmS of from Batlogg et(see al., who studied thebasic optical and electrical behavior of theproperties semiconducting comes from state. Batlogg al.,have whoalready studieddemonstrated, the optical andthe electrical behavior the semiconducting and and metallic Asetwe cubic crystal fieldofsplits the 5d band in two metallic state. As we have already demonstrated, the cubic crystal field splits the 5d band in two sub-bands. In the metallic phase, the peak in the imaginary part of the dielectric function at around sub-bands. In the metallic phase, thethe peak the imaginary partofofSm theions dielectric function at around 5.5 eV is related to the transition from 3p6inanion valence band to empty states above the 6 5 6 5.5 eVlevel. is related to the transition fromatthe 3p 4.5 anion valence band of Sm ions to the empty Fermi Additionally, a steep kink about eV stems from transitions from 4f (states H5/2 )above state 3+ to empty the Fermi level. Additionally, a steep kink at about eV stems transitions from the 4f5(6H5/2 of Sm states above the Fermi level. In the4.5 visible rangefrom of the metallic phase, a valley is) 3+ stateatof3.1 SmeV,todue empty states aboveof the Fermi level. (Drude In the visible range the metallic phase, a valley seen to the presence free electrons model). Theofgolden color of the metallic is seen at 3.1 eV, due to the presence of free electrons (Drude model). The golden color of the metallic state is due to the reflectivity drop at around 0.45 µm (Figure 3a). This valley of the reflectivity at state is wavelength due to the reflectivity drop at around 0.45 μm (Figure 3a).SmS Thisinvalley of regime, the reflectivity at shorter is related the plasma frequency of the metallic the UV with the shorter wavelength related the plasma the metallic SmStransitions in the UV[15,36–39]. regime, with the resulting sharp dip atis0.45 µm coming fromfrequency inter-bandof(bound electron) resulting sharp dip at 0.45 μm coming from inter-band (bound electron) transitions [15,36–39]. Metallic SmS can be stable on the surface of a sample and can be obtained either by polishing, Metallic SmS can be stable on the surface of a sample and can be obtained either polishing, pressurizing or by scratching the surface of a semiconducting SmS sample. Nevertheless, aby part of SmS pressurizing or by scratching surface of a semiconducting SmS sample. Nevertheless, a part of can probably remain unaltered, the in its semiconducting state, as switching can occur on a grain-to-grain SmS[15]. can Reflection probably remain unaltered, in used its semiconducting state, asxswitching can occur on a in grainbasis measurements can be to show that an amount of SmS remains unaltered, the to-grain basis [15]. can used to show that an amount SmS remains semiconducting state.Reflection As a result,measurements the reflectivity of thebesample (with reflectivity R(SmS)mixx) isofthe sum of the unaltered, the semiconducting a result, the reflectivity of the sample (with reflectivity two fractionsin(with reflectivity RM-SmSstate. and RAs S-SmS ) that simultaneously exist within the sample. In order to R (SmS)mix ) is the sum of the two fractions (with reflectivity RM-SmS and R S-SmS) that simultaneously exist calculate the reflectivity of the metallic state, we can use: RM-SmS = 1.18(R (SmS)mix − xR-SmS ) [28]. within the sample. In order to calculate the reflectivity of the metallic state, we can use: RM-SmS = 1.18(R(SmS)mix − xR-SmS) [28].


5 of 21 5 of 21

Figure 3. (a) SmS reflectivity as a function of wavelength in both metallic and semiconducting phase (SmS single singlecrystal); crystal);(b) (b)Reflectivity Reflectivity(SmS (SmSsingle singlecrystal) crystal)ofof the semiconducting phase. The final levels (SmS the semiconducting phase. The final levels of 6 7 6 7 of the transitions labelled graph, while thetransitions transitionsstart startfrom from4f4f( ( FF00); (c) Transmission the transitions areare labelled in in thethe graph, while allall ofofthe film S-SmS. S-SmS. ((a) ((a) and and (b) (b) adapted adapted from from [39] [39] and and [30], [30], respectively, respectively,with withpermission.) permission.) of a 100 nm thin film

Very recently, recently, some some investigations investigations have have shown shown the the existence existence of of aa pseudo-gap, pseudo-gap, before before the the final final Very metallic behavior behavior is is reached, reached, upon upon application application of of pressure. pressure. This metallic This effect effect has has been been studied studied in in detail detail [40], [40], by measuring the optical reflectivity under pressure, in both the middle and far infrared region of the by measuring the optical reflectivity under pressure, in both the middle and far infrared region of electromagnetic spectrum. It was observed that the black–golden transition takes place when the gap the electromagnetic spectrum. It was observed that the black–golden transition takes place when the size size getsgets almost equal to the excitonic binding energy before full gap almost equal to the excitonic binding energy beforeit itentirely entirelycloses closesto to achieve achieve the the full metallic state. As a result, this transition essentially stems from an excitonic instability. The energy metallic state. As a result, this transition essentially stems from an excitonic instability. The energy gap gap at a pressure of 0.65 measured be±4520± 20 meV, with excitonic binding energy equal at a pressure of 0.65 GPaGPa waswas measured to beto45 meV, with an an excitonic binding energy equal to to 60 meV. Upon the semiconductor–metal transition, the electrons are mixed in the 5d conduction 60 meV. Upon the semiconductor–metal transition, the electrons are mixed in the 5d conduction band bandthe with 4f of band the divalent ion,awith a transformation from the divalent to trivalent with 4f the band the of divalent ion, with transformation from the divalent to trivalent state ofstate Sm of Sm ions. As the trivalent state has a smaller ionic radius, the lattice constant reduces, providing ions. As the trivalent state has a smaller ionic radius, the lattice constant reduces, providing a furthera further lowering of the 5d band, below4fthe initial 4f state, leading tochange the valence lowering in energyinofenergy the lowest 5dlowest band, below the initial state, leading to the valence from 2+ to mainly Sm3+ [40]. change from mainly Sm 2+ 3+ mainly Sm to mainly Sm [40]. An intermediate intermediate valence valence state determined by the Fermi Fermi level in the the gap gap An state is is determined by the the exact exact position position of of the level in 6 level and especially its location in comparison with a virtual localized state, which stems from the 4f 6 and especially its location in comparison with a virtual localized state, which stems from the 4f level 2+ of the this virtual state, oneone obtains the the divalent state,state, while a trivalent state of the Sm Sm2+ground groundstate. state.Above Above this virtual state, obtains divalent while a trivalent occurs belowbelow it. When the Fermi energyenergy is pinned to this to state, obtain an intermediate or a mixedstate occurs it. When the Fermi is pinned thiswe state, we obtain an intermediate or a valence state, which could be explained by two possible mechanisms. The first one is related to a mixed-valence state, which could be explained by two possible mechanisms. The first one is related spatial mixture of valence states, in the case both divalent the to a spatial mixture of valence states, in the case both divalentand andtrivalent trivalentions ionsare are present present in in the material, providing a mixing and disorder on the atomic scale. The second interpretation comes from material, providing a mixing and disorder on the atomic scale. The second interpretation comes from a a temporal mixture, when every ion temporally fluctuates between thepossible two possible valence temporal mixture, when every Sm Sm ion temporally fluctuates between the two valence states. states. In practice, the second explanation more probable for several reasons, as the prohibitive In practice, the second explanation is moreisprobable for several reasons, such assuch the prohibitive strain strain energy of the spatial energy of the spatial mixingmixing [12,41].[12,41].

Materials 2017, 10, 953

6 of 21

2.3. Resistivity Drop and Structural Properties SmS single crystals revert to the semiconducting state upon the release of the pressure, used for inducing the metallic state [12]. SmS thin films do not necessarily show this reversible behavior, as tensile pressure or annealing in vacuum is necessary to provide the back switching to the initial semiconducting state. Recently, Imura et al. [42] studied the electronic structure of Sm1−x Yx S (x = 0, 0.03, 0.12, or 0.32) single crystals by using ARPES in order to investigate the black–golden phase transition. By substitution of Sm by Y, the gap closes and the metallic state can be stabilized at room temperature (Figure 4a). The stabilization of the metallic state at ambient pressure occurs around a critical Y concentration of x = 0.2, due to the presence of intrinsic stress. The closing of the gap is visible in the ARPES spectra (Figure 4b) for increasing Y concentration. The solid and broken lines indicate the multiplet structure of the Sm 4f5 states (6 H, 6 F, 6 P), where a discrimination is made between bulk and surface states. A similar effect can be obtained by using gadolinium instead of yttrium as an alloying element. Sharenkova et al. [43], studied Sm1−x Gdx S single crystals prepared by directed crystallization from the melt. Based on XRD measurements before and after the valence transition, it was substantiated that the decrease of scattering regions—which stem from misfit dislocations—is a crucial factor influencing the transition. The decrease of the size of the scattering regions is a result rather than a parameter, influencing the final pressure-induced transition in10, SmS-based materials [43]. Materials 2017, 953 7 of 21

Figure 4. 4. (a) (a) Yttrium Yttrium concentration concentration dependence dependence of of the the lattice lattice constant constant of ofthe theSmS-based SmS-basedsingle singlecrystal; crystal; (b) The density of states at the Γ point of the Brillouin zone of the SmS-based SmS-based material. The photon energies energies and and concentrations concentrations of of YY are are placed placed on on the the left left and and the the right right side sideof ofthe thegraph, graph,respectively. respectively. (Reprinted with permission from [42].)

X-ray absorption near edge (XANES) are shown in thus Figure where two Both the transition and the spectroscopy resistivity change uponresults the transition can be 5a, influenced by XANES white lines are observed at about 6711.7 eV and 6719.3 eV, for divalent and trivalent Sm, substituting Sm with other elements. When substituting a large part of Sm by Eu, the transition is respectively. At 4.5 K and 0 GPa (black triangles), only the divalent state of Sm was observed. By impaired in (Sm,Eu)S. Substitution of Sm by Eu (ionic radius of divalent Eu: 0.131 nm) leads to a shift 3+ increasing the pressure 1.85 corresponding totransition the golden phase, values. the white line for Smis of the pressure thresholdtofor theGPa, semiconductor–metal to higher This behavior dominates, although certain fraction of the Sm ions remains in the divalent which accompanied by a shifta of the lowest 5d band to higher energies. A different behaviorstate, is observed demonstrates the mixed-valence character [26,44]. In Figure 5b, we observe a similar diagram the when Yb (ionic radii of divalent Yb: 0.116 nm) is used, as even a large amount of Yb inducesinonly case of the Sm L3 XANES spectrum of Sm 0.55Y0.45S, at several temperatures [45]. At lower a relatively small reduction of the resistivity for the alloyed system, as the position of the 5d band concentrations of Y,unaltered. the Sm2+ peak is stronger shown), about peakelement, intensityhaving in the 2+ asequal remains relatively Another case is(not found in theand usehas of Ca alloying 3+ peak becomes dominant. case of Sm 0.67Y0.33S. Increasing the temperature in both cases, the Sm 2+ an ionic radius (0.114 nm) comparable to Sm (0.109 nm). This differs from the previous alloys, as Specifically, the case Sm0.55Y0.45band S trivalent becomes abundant. Forstructure these two concentrations, CaS presentsin neither 5d of conduction nor 4fSm levels. Both the electronic and the size effect there is no sharp increase of lattice parameter, as the material acquires its semiconducting state, which can however influence the switching and resistivity behavior of SmS [25]. In this case, the resistivity is anticipated in the case of SmS. Possibly, the increase in temperature leads to an increase of Sm valence, with associated decrease of the lattice parameter, with presumably stronger 4f electron delocalization [45]. Several competing factors determine whether the semiconductor–metal transition is continuous or discontinuous [26]. The lattice contraction during the transition from Sm2+ to Sm3+ is counteracted

Materials 2017, 10, 953

7 of 21

behaves similarly to the usage of EuS: for lower concentrations of Ca, a first order abrupt resistivity drop occurs, while for higher Ca concentration the system shows a continuous transition from the high to the low resistivity. For alloyed SmS-based thin films, the mechanism behind the transition is explained as follows. In SmS, the system possesses its semiconducting phase under normal conditions. The substitution of additional elements into the structure of SmS such as Y or Eu promotes charge carriers into the 5d conduction band, with the lattice constant of the semiconducting SmS decreasing and approaching metallic values. After that, the transition takes place without any additional contribution, such as high pressure. Subsequently, as we can see in Figure 4a, the system collapses when the fraction of additional element in the structure reaches the critical concentration xc . Nonetheless, the multiplet structure of the 4f6 ground state of Sm2+ ion fully touches the Fermi level only at concentrations higher than xc . Overall, the transition in this kind of system takes place upon reaching the concentration xc , where the band gap between 4f and 5d vanishes [14,43]. The effect of Sm substitution on the SmS lattice parameter is dependent on the alloying element, and can be divided in three cases. When using Gd or Nd, as alloying elements, the lattice parameter shows a discontinuous decrease as a function of increasing Gd or Nd concentration. In the case of Yb and Ca, no discontinuous behavior is observed, while the usage of Eu (EuS shows almost the same lattice constant (0.596 nm) with S-SmS) does not lead to any change in lattice parameter over the whole range of concentrations. X-ray absorption near edge spectroscopy (XANES) results are shown in Figure 5a, where two XANES white lines are observed at about 6711.7 eV and 6719.3 eV, for divalent and trivalent Sm, respectively. At 4.5 K and 0 GPa (black triangles), only the divalent state of Sm was observed. By increasing the pressure to 1.85 GPa, corresponding to the golden phase, the white line for Sm3+ dominates, although a certain fraction of the Sm ions remains in the divalent state, which demonstrates the mixed-valence character [26,44]. In Figure 5b, we observe a similar diagram in the case of the Sm L3 XANES spectrum of Sm0.55 Y0.45 S, at several temperatures [45]. At lower concentrations of Y, the Sm2+ peak is stronger (not shown), and has about equal peak intensity in the case of Sm0.67 Y0.33 S. Increasing the temperature in both cases, the Sm3+ peak becomes dominant. Specifically, in the case of Sm0.55 Y0.45 S trivalent Sm becomes abundant. For these two concentrations, there is no sharp increase of lattice parameter, as the material acquires its semiconducting state, which is anticipated in the case of SmS. Possibly, the increase in temperature leads to an increase of Sm valence, with associated decrease of the lattice parameter, with presumably stronger 4f electron delocalization [45]. Several competing factors determine whether the semiconductor–metal transition is continuous or discontinuous [26]. The lattice contraction during the transition from Sm2+ to Sm3+ is counteracted by e.g., the higher bulk modulus at higher pressure (50 GPa at zero pressure, [28]) and the different density of states between the 4f and 5d bands. In addition, the 6s band has a crucial influence on the transition. If there is a strong hybridization between 5d and 6s states at the bottom of the conduction band, the transition will probably be continuous, due to the relatively low density of states at the Fermi level. If there is only 5d contribution to the bottom part of the conduction band, then a discontinuous or first-order transition is likely ([25] and papers therein). By using X-ray photoelectron spectroscopy (XPS), the thermally induced transition between the two states, from the golden to the black one can also be observed [46]. XPS measurements have shown that both the semiconductor–metal transition and the oxidation process (a fraction of Sm oxide appearing after annealing) in SmS are responsible for the final change in color [47]. In another investigation, the changes in resistance were studied after pulsed laser irradiation at 308 nm, demonstrating that a fine modulation of the resistance of the black state of SmS can be achieved at lower fluences in comparison with the laser ablation fluence, without significant destruction of the films. The reason of this structural film modification is related to surface tension forces and thermally induced stresses [48]. In comparison with other work [49], these thin films presented less structural damage during transition.


8 of 21 8 of 21

Figure5. 5.(a) (a)X-ray X-rayabsorption absorption spectra spectra at 4.5 Black triangles Figure 4.5 K K of of the the Sm Sm2+2+L3 L3transition transitionofofSmS. SmS. Black triangles correspondtotoa apressure pressureof of00GPa, GPa, while while the red GPa; (b)(b) XANES correspond red dots dots are areupon uponapplication applicationofof1.85 1.85 GPa; XANES spectra Sm 0.55 Y 0.45S thin films at several temperatures, without application of pressure. (Reprinted spectra ofof Sm Y S thin films at several temperatures, without application of pressure. (Reprinted 0.55 0.45 with permissionfrom from[26] [26]and and[45], [45],respectively.) respectively.) with permission

2.4. Magnetic Properties 2.4. Magnetic Properties SmS presents a pseudo-gap in the metallic state and shows paramagnetic behavior, while it is SmS presents a pseudo-gap in the metallic state and shows paramagnetic behavior, while it is non-magnetic in the semiconducting state. The mean valence v (showing the number of electrons per non-magnetic in the semiconducting state. The mean valence v (showing the number of electrons per Sm ion) and conductivity σ of SmS single crystal have a continuous monotonic increase with pressure Sm ion) and conductivity σ of SmS single crystal have a continuous monotonic increase with pressure at high T, as can be observed in Figure 6a,b. At low temperature, there is a critical pressure 1.8 GPa at for high the T, as magnetic can be observed in Figure low temperature, thereatis alower criticalpressures pressure 1.8 transition from6a,b. theAt paramagnetic state to GPa thefor theantiferromagnetically magnetic transitionordered from thestate paramagnetic state at This lower to measuring the antiferromagnetically at higher pressure. is pressures revealed by the magnetic ordered state at higher pressure. This is revealed by measuring the magnetic susceptibility susceptibility (Figure 6c, yellow shapes). Consequently, a quantum phase transition occurs (Figure (at low 6c, yellow shapes). Consequently, a quantum phase transition occurs (at low temperature) at this pressure temperature) at this pressure (Pc in the graph), which originates from the large variation in the (Pcnumber in the of graph), which from the large 6d variation inreferences the number of carriers n atthus T=2 carriers n at Toriginates = 2 K, as shown in Figure ([50] and therein). There are K,two as shown Figure 6din([50] references therein).transition There are two kinds of transitions kinds ofintransitions SmS,and a semiconductor–metal andthus a magnetic transition from the in SmS, a semiconductor–metal transition and a magnetic from the paramagnetic state to the paramagnetic state to the antiferromagnetically ordered transition state. In Figure 6c, T0 and TN symbolize antiferromagnetically ordered state. the temperature of the opening of a pseudo-gap and the Néel temperature, respectively. star (tricritical point) marks a first order In Figure 6c, T0 and TNThe symbolize the temperature of the the separation opening ofbetween a pseudo-gap and the transition, indicated by the double line, and a second order transition (single line). Overall, there Néel temperature, respectively. The star (tricritical point) marks the separation between a first are order two important aspects about the magnetic metallic SmS. Firstly, the Curie law provides transition, indicated by the double line, andbehavior a secondoforder transition (single line). Overall, there are a description the localized moments at high temperature, withSmS. a pressure state. a two important of aspects about the magnetic behavior of metallic Firstly,dependent the Curieground law provides In the case the pressure is lower than the critical one (1.8 GPa) and the temperature lower than T0, In description of the localized moments at high temperature, with a pressure dependent ground state. these localized moments disappear due to spin-singlet and exciton-like bound states. Secondly, the case the pressure is lower than the critical one (1.8 GPa) and the temperature lower than T0 , by these contrast, at a pressure greater than the critical pressure, the localized moments form an localized moments disappear due to spin-singlet and exciton-like bound states. Secondly, by contrast, antiferromagnetic state, below TN promoting the RKKY (Ruderman-Kittel-Kasuya-Yosida) at a pressure greater than the critical pressure, the localized moments form an antiferromagnetic state, interaction. This interaction stems from the coupling mechanism in a metal between its localized below TN promoting the RKKY (Ruderman-Kittel-Kasuya-Yosida) interaction. This interaction stems either f or d shell electron spins and nuclear magnetic moments with electrons of the conduction band from the coupling mechanism in a metal between its localized either f or d shell electron spins and via heavy quasiparticles. This means that there is a transition from a real bound state to a Kondo nuclear magnetic moments with electrons of the conduction band via heavy quasiparticles. This means virtual bound state at this specific critical pressure [51]. It should be mentioned that this behavior is that there a transition from a real boundantiferromagnetic state to a Kondo ordering virtual bound state at thisformation specific critical possiblyis related to a competition between and pseudo-gap due pressure [51]. It should be mentioned that this behavior is possibly related to a competition between to screening [40,51]. antiferromagnetic ordering and pseudo-gap formation due to screening [40,51].


9 of 21 9 of 21

Figure 6. (a) Figure 6. (a) Mean Mean valence valence (already (already well well above above valence valence state state 2) 2) and and (b) (b) conductivity conductivity of of SmS SmS at at the the specific pressure interval between approximately 0.5 GPa and 2.4 GPa. Different shapes of (a) specific pressure interval between approximately 0.5 GPa and 2.4 GPa. Different shapes of (a) represent represent valuesfrom coming from different works, in the original paper of Figure 6. Black line serves values coming different works, cited in cited the original paper of Figure 6. Black line serves as a as a guide toeye; the (c) eye;P-T (c)phase P-T phase diagram the metallic PM: paramagnetic regime, guide to the diagram of theofmetallic phase,phase, wherewhere PM: paramagnetic regime, AFM: AFM: antiferromagnetic and DM: diamagnetic regime; (d) Evolution of theofnumber of free antiferromagnetic regime regime and DM: diamagnetic regime; (d) Evolution of the number free carriers (n) carriers (n) at T = 2 K. (Reprinted with permission from [50].) at T = 2 K. (Reprinted with permission from [50].)

The most important reason to study the magnetic properties of SmS is the instability between The most important reason to study the magnetic properties of SmS is the instability between the magnetic states at the quantum critical point (QCP, at T = 10 K and P = 1.8 GPa). In this regime, the magnetic states at the quantum critical point (QCP, at T = 10 K and P = 1.8 GPa). In this regime, phenomena such as unconventional superconductivity and the Kondo effect appear [13]. As phenomena such as unconventional superconductivity and the Kondo effect appear [13]. As mentioned, mentioned, the Sm divalent state is non-magnetic, since the total magnetic moment is zero (J = 0), as the Sm divalent state is non-magnetic, since the total magnetic moment is zero (J = 0), as L = S = 3, while L = 3+ S = 3, while Sm3+ shows magnetic behavior. Barla et al. [52] studied the SmS magnetic properties Sm shows magnetic behavior. Barla et al. [52] studied the SmS magnetic properties by measuring by measuring Nuclear Forward Scattering (NFS) spectra at various temperatures and pressures of Nuclear Forward Scattering (NFS) spectra at various temperatures and pressures of metallic SmS metallic SmS (Figure 7). By using this technique, the quantum beat pattern was obtained from the (Figure 7). By using this technique, the quantum beat pattern was obtained from the combined action combined action of magnetic dipole and electric quadrupole interactions on the nuclear level of the of magnetic dipole and electric quadrupole interactions on the nuclear level of the isotope 149 Sm [52]. 149 isotope Sm [52]. This behavior seems to exist only at low temperatures. The reason that no effect This behavior seems to exist only at low temperatures. The reason that no effect was observed at was observed at higher temperatures (for a variety of pressures) is related to a common behavior of higher temperatures (for a variety of pressures) is related to a common behavior of unsplit nuclear unsplit nuclear levels for Sm ions in the absence of magnetic order within a cubic symmetry levels for Sm ions in the absence of magnetic order within a cubic symmetry environment [52]. It was environment [52]. It was shown that only part of the Sm atoms, corresponding to 72%, at 2.35 GPa shown that only part of the Sm atoms, corresponding to 72%, at 2.35 GPa and 3 K showed a magnetic and 3 K showed a magnetic order, the remainder being paramagnetic. This amount (28%, order, the remainder being paramagnetic. This amount (28%, paramagnetic) was transferred to its final paramagnetic) was transferred to its final magnetic state at 3 GPa. magnetic state at 3 GPa. At temperatures and pressures lower than T0 and Pc, the ground state of metallic SmS is At temperatures and pressures lower than T0 and Pc , the ground state of metallic SmS is diamagnetic, as the magnetic susceptibility does not demonstrate any paramagnetic Curie–Weiss diamagnetic, as the magnetic susceptibility does not demonstrate any paramagnetic Curie–Weiss divergence [53]. Several studies [13,50] have reported the temperature dependence of the thermal divergence [53]. Several studies [13,50] have reported the temperature dependence of the thermal expansion coefficient α(T) of metallic SmS. All these measurements took place by using SmS single expansion coefficient α(T) of metallic SmS. All these measurements took place by using SmS single crystals. NFS and thermal expansion experiments show a tricritical point (Tc), as we see in Figure 6c crystals. NFS and thermal expansion experiments show a tricritical point (Tc ), as we see in Figure 6c (star symbol), where the separation of the first order Néel transition from the second order transition (star symbol), where the separation of the first order Néel transition from the second order transition takes place. By combining all these results, a P-T diagram can be obtained similar to Figure 6c, takes place. By combining all these results, a P-T diagram can be obtained similar to Figure 6c, consisting of the first and second order lines due to the Néel temperature, and the crossover line stemming from T0.

Materials 2017, 10, 953

10 of 21

consisting of the first and second order lines due to the Néel temperature, and the crossover line stemming from T0 . Materials 2017, 10, 953 10 of 21

149 Sm NFS spectra of SmS at different Figure Figure7. 7. 149Sm NFS spectra of SmS at different temperatures and pressures. pressures. Dots Dots are are the the experimental experimental results, while the full lines are fits. (Reprinted with permission from [52].) results, while the full lines are fits. (Reprinted with permission from [52].)

Thereare aretwo twomain maincharacteristic characteristicpeaks peaksin in the the temperature temperaturedependence dependenceof ofthe the thermal thermal expansion expansion There coefficient. A shallow minimum and positive sharp peaks are seen in Figure 8a, which are relatedto to coefficient. A shallow minimum and positive sharp peaks are seen in Figure 8a, which are related and the the Néel Néel temperature, temperature, respectively. respectively. The The positive positive peak peak at at pressures pressures above above the the second second critical critical TT00 and pressure can be understood from the inset graph ΔL/L(T) as well, where we see a strong change at pressure can be understood from the inset graph ∆L/L(T) as well, where we see a strong change at these pressures [13]. In Figure 8b, the evolution of the thermal expansion coefficient for a pressure these pressures [13]. In Figure 8b, the evolution of the thermal expansion coefficient for a pressure interval between between 16.2 16.2 and shallow minimum at 10 interval and 19.1 19.1 kbar kbar(1.62 (1.62and and1.91 1.91GPa) GPa)isisshown. shown.The The shallow minimum at K 10for K low pressure changes to a sharp positive peak upon increasing the pressure. This describes the mixing for low pressure changes to a sharp positive peak upon increasing the pressure. This describes the of the two magnetic phases,phases, which which consists of a broad minimum and a and positive peak peak stemming from mixing of the two magnetic consists of a broad minimum a positive stemming the paramagnetic and and antiferromagnetic states, respectively. The The temperature dependence of the from the paramagnetic antiferromagnetic states, respectively. temperature dependence of compressibility shows a similar behavior: broader peaks at lower pressure and narrower peaks close the compressibility shows a similar behavior: broader peaks at lower pressure and narrower peaks to TNto , which is consistent withwith the behavior of α(T). Measuring magnetic susceptibility χ and the close TN , which is consistent the behavior of α(T). Measuring magnetic susceptibility χ and H is another way to obtain T N [36,54]. In pressure pressure dependence of the Hall coefficient R the pressure dependence of the Hall coefficient RH is another way to obtain TN [36,54]. In pressure dependentmeasurements measurementsof ofRRHH of of the the metallic metallic state state of of SmS, SmS, there there is is aa transition transition from from values values above above 00 dependent to values values below below 00 at at aapressure pressureof ofabout about1.75 1.75GPa GPaat atconstant constanttemperature temperature 22 K, K,which whichdescribes describes the the to magnetic transition. Exploiting all these kinds of anomalies, a P-T graph similar to that of the Figure 6c magnetic transition. Exploiting all these kinds of anomalies, a P-T graph similar to that of the Figure 6c can be made. can be made. In the the previous previous part, part, we we analyzed analyzed results results from from optoelectronic optoelectronic and and magnetic magnetic properties properties of of SmS. SmS. In In this part, we will discuss more exotic properties of this material, such as electrical oscillations and In this part, we will discuss more exotic properties of this material, such as electrical oscillations and Kondoinsulator insulatorbehavior. behavior. A very recent investigation electrical oscillations in SmS was Kondo A very recent investigation about about electrical oscillations in SmS was reported reported by Takahashi et al. [55]. Another way to induce the semiconducting–metallic transition, by Takahashi et al. [55]. Another way to induce the semiconducting–metallic transition, apart from apart pressure, from using pressure, is by applying DC electric order of a few kV/cm), inducing using is by applying a DC electricafield (on the field order(on of athe few kV/cm), inducing a collective 3 V/cm), 3 V/cm), a collective motioncausing of charges causing an electrical[55]. oscillation At high (4 × 10 motion of charges an electrical oscillation At high[55]. electric fieldselectric (4 × 10fields nonlinear nonlinear conduction areElectrical observed.oscillations Electrical oscillations are aneffect, additional conduction phenomenaphenomena are observed. are an additional whicheffect, takeswhich place takes place during the semiconducting–metallic transition and it has been observed in such several during the semiconducting–metallic transition and it has been observed in several materials, as materials, such as correlated insulators [56,57]. In the case of the semiconducting SmS, a persistent correlated insulators [56,57]. In the case of the semiconducting SmS, a persistent electrical oscillation electrical oscillationby hasapplying been observed byexternal applyingvoltage constant external voltage the (DC). In general, the fhas been observed constant (DC). In general, f-electron systems electron systemscorrelation show prominent correlation behavior, is the main reason of the show prominent behavior, which is the main reasonwhich of the aforementioned phenomenon, aforementioned phenomenon, which is very similar to the tunneling mechanism and negative differential resistance in narrow gap semiconductors (Zener effect) [55].

Materials 2017, 10, 953

11 of 21

which is very similar to the tunneling mechanism and negative differential resistance in narrow gap Materials 2017, 10, 953 11 of 21 semiconductors (Zener effect) [55].

Figure Figure 8. 8. (a) (a) Temperature Temperature dependence dependence of of the the thermal thermal expansion expansion coefficient coefficient at at pressures pressures up up to to 21.2 21.2 kbar kbar (2.12 GPa), with T 0 and TN corresponding to shallow negative and sharp positive peaks, respectively (2.12 GPa), with T0 and TN corresponding to shallow negative and sharp positive peaks, respectively and and aa ΔL/L(T) ∆L/L(T)graph graphin inthe theinset. inset.Data Datahave havebeen beenshifted shifted by by the the authors authors for for clarity; clarity; (b) (b) A A similar similar graph graph at between 16.2 16.2 and and 19.1 19.1 kbar kbar (1.62–1.91 (1.62–1.91 GPa). GPa). The at pressures pressures between The area area which which is is created created by by the the two two vertical vertical dashed linesin inthe theinset insetdemonstrates demonstrates mixed phase regime. (Reprinted permission from dashed lines thethe mixed phase regime. (Reprinted withwith permission from [13].) [13].)

It is still a point of discussion whether a Kondo insulating behavior is present in SmS. A Kondo It is still a point of discussion whether a Kondo insulating behavior is present in SmS. A Kondo insulator is a material for which the electrical resistivity dramatically increases—a narrow band gap insulator is a material for which the electrical resistivity dramatically increases—a narrow band gap opens—at low temperatures. This is a result of the hybridization of localized electrons with electrons opens—at low temperatures. This is a result of the hybridization of localized electrons with electrons of the conduction band. Previously, we have noted the existence of a mixed valence state and a of the conduction band. Previously, we have noted the existence of a mixed valence state and a pseudo-gap in SmS. The excitonic instability is a factor that determines the pseudo-gap, but, in another pseudo-gap in SmS. The excitonic instability is a factor that determines the pseudo-gap, but, in work, the hybridization of the 4f band into 4f5/2 and 4f7/2 has also been mentioned as an important another work, the hybridization of the 4f band into 4f5/2 and 4f7/2 has also been mentioned as an parameter. The 4f5/2 band is placed quite close to the Fermi energy, whilst the 4f7/2 level is located important parameter. The 4f5/2 band is placed quite close to the Fermi energy, whilst the 4f7/2 level is about ±0.17 eV away from the Fermi energy [51]. Kondo behavior stems from strongly correlated located about ±0.17 eV away from the Fermi energy [51]. Kondo behavior stems from strongly 4f systems, so SmS is a possible candidate, as described by Li et al. [58]. However, Kang et al. [51] correlated 4f systems, so SmS is a possible candidate, as described by Li et al. [58]. However, Kang et proposed a different model for SmS instead of that of a conventional Kondo insulator. In that case, al. [51] proposed a different model for SmS instead of that of a conventional Kondo insulator. In that the behavior of SmS has been analyzed by using dynamical mean-field theory (DMFT), properly case, the behavior of SmS has been analyzed by using dynamical mean-field theory (DMFT), properly describing the insulating ground state of SmS and the topological properties of the SmS metallic state describing the insulating ground state of SmS and the topological properties of the SmS metallic state of strongly correlated 4f electrons. In order to study the topological properties of metallic SmS, a surface of strongly correlated 4f electrons. In order to study the topological properties of metallic SmS, a band structure has been calculated (not shown). A tiny gap was found, which demonstrates that surface band structure has been calculated (not shown). A tiny gap was found, which demonstrates Rashba spin-polarized surface states (i.e., a momentum dependent splitting of spin states) play a role that Rashba spin-polarized surface states (i.e., a momentum dependent splitting of spin states) play instead of topologically protected Kondo states with a Dirac cone in the band structure. Nonetheless, a role instead of topologically protected Kondo states with a Dirac cone in the band structure. elsewhere [59], Sm0.75 La0.25 S was found to have an excitonic insulating gap equal to 1 meV, describing Nonetheless, elsewhere [59], Sm0.75La0.25S was found to have an excitonic insulating gap equal to 1 it as a possible Kondo insulator. Since then, SmS has been characterized as a Kondo semimetal, instead meV, describing it as a possible Kondo insulator. Since then, SmS has been characterized as a Kondo of a Kondo insulator, like the very similar compound SmB6 (showing an intermediate valence at semimetal, instead of a Kondo insulator, like the very similar compound SmB6 (showing an ambient pressure) [23,60]. intermediate valence at ambient pressure) [23,60]. 3. Thin Film and Bulk SmS Preparation Techniques 3. Thin Film and Bulk SmS Preparation Techniques 3.1. Electron Beam Evaporation—Reactive Evaporation 3.1. Electron Beam Evaporation—Reactive Evaporation Electron beam evaporation (EBE) is a relatively cheap and flexible method for the deposition of a Electron evaporation (EBE)Apart is a relatively flexible method for deposition the deposition of wide range ofbeam compound thin films. from the cheap choiceand of evaporation source, speed aand wide range oftemperature, compound thin Apart from thestoichiometry choice of evaporation source, deposition speed substrate the films. film properties and can be influenced to some extent and substrate temperature, the film stoichiometry can influenced to some by reactive evaporation. In that case,properties a reactive and gas is introduced in thebevacuum system, withextent a low by reactive evaporation. In that case, a reactive gas is introduced in the vacuum system, with a low pressure, limited by the maximum working pressure of the e-beam filament. The e-beam deposition of pressure, limited by the maximum working pressure of the e-beam filament. The e-beam deposition of SmS is often started from a Sm metal source, accompanied by an H2S flow. Rogers et al. described the deposition of homogeneous thin films by using e-beam evaporation and appropriately balancing the Sm evaporation rate and the H2S partial pressure, to arrive at the desired stoichiometry. A continuous semiconductor–metal transition was observed from optical measurements, but

obtained [15,61]. Another similar investigation based on e-beam evaporation comes from Hickey et al., who used several types of substrates, such as PMMA, quartz and soda lime glass. The main characteristic was the presence of a relatively high level of impurities, such as oxygen [12,41]. Materials 2017, 10, 953 3.2. Sputtering

12 of 21

Sputtering, in both RF and DC operation, has been used for the deposition of SmS [62,63]. Given SmS from a Sm metal source, accompanied by DC an Hmagnetron et al. described that isa often SmS started target is difficult to obtain, one can combine sputtering using a the Sm 2 S flow. Rogers deposition of homogeneous thin films using e-beam and appropriately the (metallic) target and an additional RFby magnetron withevaporation a chalcogenide (Sm2S3) target balancing to eventually obtain SmS. The structural of the resulting determined usingAXRD. Three Sm evaporation rate and the Hproperties to arrivefilms at thewere desired stoichiometry. continuous 2 S partial pressure, possible phases couldtransition be obtained on theoptical ratio Pmeasurements, rf/Pdc (see figure 9).simultaneously For example, the semiconductor–metal wasdepending observed from but a metallic phase transition was obtained a ratioplace equalon to 2, a quasi-metallic state (intermediate state)published for a ratio discontinuous wasfor taking a grain to grain basis, similar to results equal to 3.2 and the final semiconducting was obtained for astructure power ratio of 5.46. elsewhere [15,61]. The XRD peaks pointedphase at a randomly oriented in comparison with thin anotherusing investigation [61], by properly changing the substrate the power of films In deposited sputtering techniques, where strongly textured thintemperature films can beand obtained [15,61]. the Sm and Sm2investigation S3 RF sputterbased targets thus their deposition different phases of Smseveral xS (x is Another similar on(and e-beam evaporation comes rates), from Hickey et al., who used the ratio of Sm to S)such could deposited, without anylime traceglass. of SmO. the most intriguing types of substrates, as be PMMA, quartz and soda ThePossibly, main characteristic was the result of of this work is that forlevel x > 1, metallic such phaseasofoxygen SmS was obtained as-deposited. In all cases presence a relatively high ofthe impurities, [12,41]. of deposition power and substrate temperature, the most prominent diffraction peak was that of 3.2. Sputtering (200). Nonetheless, there was a splitting of this peak into two peaks at higher deposition temperatures, atand 275 DC °C, operation, indicating has the been simultaneous presence of theofsemiconducting and Sputtering,especially in both RF used for the deposition SmS [62,63]. Given metallic phase. was mainly observed (away fromDC x = magnetron 1) in the case of x = 3.8using and xa =Sm 2.3. When x that a SmS targetThis is difficult to obtain, one can combine sputtering (metallic) decreased below one, new phases were created and the sulfur-rich phases Sm 3 S 4 and Sm 2 S3 SmS. were target and an additional RF magnetron with a chalcogenide (Sm2 S3 ) target to eventually obtain formed. Valuesproperties of x closeof to the oneresulting yielded afilms semiconducting SmS with corresponding lattice constant. The structural were determined using XRD. Three possible phases Generally, it is evident that one of the advantages of using a co-sputtering technique with two targets could be obtained depending on the ratio Prf /Pdc (see Figure 9). For example, the metallic phase was is that by adjusting the parameters of sputtering there is a possibility to manipulate the dominating obtained for a ratio equal to 2, a quasi-metallic state (intermediate state) for a ratio equal to 3.2 and the phase, semiconducting or metallic, in as-deposited thin films [61–63]. final semiconducting phase was obtained for a power ratio of 5.46.

Figure9.9.XRD XRD patterns demonstrating the possible three possible states (semiconducting–intermediate– Figure patterns demonstrating the three states (semiconducting–intermediate–metallic) metallic) and awith pattern the standard diffraction SmS. (Adapted fromwith [63],permission.) with permission.) and a pattern the with standard diffraction peaks ofpeaks SmS.of(Adapted from [63],

In another investigation [61], by properly changing the substrate temperature and the power of the Sm and Sm2 S3 RF sputter targets (and thus their deposition rates), different phases of Smx S (x is the ratio of Sm to S) could be deposited, without any trace of SmO. Possibly, the most intriguing result of this work is that for x > 1, the metallic phase of SmS was obtained as-deposited. In all cases of deposition power and substrate temperature, the most prominent diffraction peak was that of (200). Nonetheless, there was a splitting of this peak into two peaks at higher deposition temperatures, especially at 275 ◦ C, indicating the simultaneous presence of the semiconducting and metallic phase. This was mainly observed (away from x = 1) in the case of x = 3.8 and x = 2.3. When x decreased below one, new phases were created and the sulfur-rich phases Sm3 S4 and Sm2 S3 were formed. Values of x

Materials 2017, 10, 953

13 of 21

close to one yielded a semiconducting SmS with corresponding lattice constant. Generally, it is evident that one of the advantages of using a co-sputtering technique with two targets is that by adjusting the parameters of sputtering there is a possibility to manipulate the dominating phase, semiconducting or Materials 2017, 10, 953 13 of 21 metallic, in as-deposited thin films [61–63].

3.3. Pulsed Laser Deposition 3.3. Pulsed Laser Deposition For pulsed laser deposition (PLD), the energetic species in the condensing flux can provide stress For pulsed laser deposition (PLD), the energetic species in the condensing flux can provide stress modification during the thin film growth. Starting from a SmS target, Zenkevich et al. manufactured modification during the thin film growth. Starting from a SmS target, Zenkevich et al. manufactured highly oriented semiconducting SmS on Si (100) at elevated substrate temperature, between 600 and highly oriented semiconducting SmS on Si (100) at elevated substrate temperature, between 600 and 1000 K [64]. In Figure 10, we see that there are no other main reflections apart from (200), with a 1000 K [64]. In Figure 10, we see that there are no other main reflections apart from (200), with corresponding lattice parameter a = 0.569 nm and an estimated grain size of 8 nm (blue line). a corresponding lattice parameter a = 0.569 nm and an estimated grain size of 8 nm (blue line). Nevertheless, the grain size, as calculated from XRD peak broadening, could be misinterpreted; the Nevertheless, the grain size, as calculated from XRD peak broadening, could be misinterpreted; the presence of stress would also increase the FWHM of the XRD peaks. An increase in the annealing presence of stress would also increase the FWHM of the XRD peaks. An increase in the annealing temperature shifted the XRD peaks to smaller angles. In the case of film annealed at T = 900 K, a lattice temperature shifted the XRD peaks to smaller angles. In the case of film annealed at T = 900 K, a lattice parameter a = 0.593 nm was obtained, corresponding to the semiconducting phase of SmS. parameter a = 0.593 nm was obtained, corresponding to the semiconducting phase of SmS.

Figure X-raypatterns patternsofof SmS films deposited PLD.blue Thecurve blueshows curvethe shows the asFigure 10. X-ray SmS thinthin films deposited usingusing PLD. The as-deposited SmS at room temperature, which presents a metallic behavior withbehavior low intensity the intensity Bragg reflection. deposited SmS at room temperature, which presents a metallic withoflow of the This behavior wasThis alsobehavior supported by also electrical and optical measurements. Themeasurements. red curve represents the Bragg reflection. was supported by electrical and optical The red Braggrepresents reflection of film at 900 K, in vacuum. The the The shape of the (200)the peak for curve theannealed Bragg reflection of annealed film at 900inset K, indetails vacuum. inset details shape thethe two cases andfor forthe an two annealing at 800 (grey dotted line). with line). permission from [64].) of (200) peak cases and forKan annealing at 800(Reprinted K (grey dotted (Reprinted with permission from [64].)

3.4. MOCVD of SmS 3.4. MOCVD of SmS Volodin et al. [65] described Molecular Organometallic Chemical Vapor Deposition (MOCVD) Volodin et al.SmS. [65] described Molecular Vapor Deposition (MOCVD) of polycrystalline In MOCVD one of theOrganometallic most importantChemical parameters is the choice of precursors. of polycrystalline SmS. In MOCVD one of the most important parameters is the choice of precursors. The following initial substances have been used in the case of SmS. Diethyldithiocarbamat (DTC) of The following initial substances havewhich been additionally used in the case of SmS. Diethyldithiocarbamat of samarium (dtc3Sm), dtc3Sm phen, contains phenantraline ligands and(DTC) dtc3Sm samarium dtc3Sm phen, whichligands. additionally contains phenantraline and in dtc3Sm bipy, which(dtc3Sm), additionally contains pyridil The MOCVD synthesis of SmS ligands is described detail bipy, which additionally contains ligands. The synthesis of SmS is described in detail by Stern et al. [60]. The best resultspyridil were obtained withMOCVD dtc3Sm bipy, providing maximum growth rate, by Stern et al. [60]. The bestand results were obtained with dtc3Sm growth polycrystallinity, continuity uniformity of the SmS thin films. bipy, Figureproviding 11 shows maximum the behavior of the rate, polycrystallinity, andtemperature uniformity offor the SmS thin films.used. Figure 11 shows behavior growth velocity versuscontinuity the substrate four precursors Using higherthe amounts of of the growth velocity versus the substrate temperature for four precursors used. Using higher ligands led to an acceleration of growth (lines 1 and 2). As anticipated, the growth rate increased at amounts of ligands led to an acceleration ofsee theingrowth and 2). anticipated, growth higher temperature. Nonetheless, as we can lines 3(lines and 4,1there wasAs a lower growththe rate when rate increased at higher temperature. Nonetheless, as we can see in lines 3 and 4, there was a lower using a waterless ambient (line 3) compared to one containing water (line 4), which was related to the growth rate using a waterless 3) compared to one containing (line 4), which purity of thewhen dtc3Sm used. The finalambient structure(line of the SmS films strongly dependedwater on parameters such was related to the purity of thematerial, dtc3Sm as used. finalprecursors. structure of SmS films depended as temperature and substrate wellThe as the Inthe general, all ofstrongly these as-deposited on parameters such as temperature and substrate material, as well as the precursors. In general, all of these as-deposited MOCVD SmS thin films showed lattice parameters close to that of the metallic (or intermediate) state of about 0.57 nm.

Materials 2017, 10, 953

14 of 21

MOCVD SmS Materials 2017, 10,thin 953 films showed lattice parameters close to that of the metallic (or intermediate) 14state of 21 of about 0.57 nm.

Figure 11. Dependence of MOCVD growth velocity of SmS for different substrate temperatures and Figure 11. Dependence of MOCVD growth velocity of SmS for different substrate temperatures and the used precursors: 1, dtc3Sm bipy; 2, dtc3Sm phen; 3, dtc3Sm (without water ambient); and 4, dtc3Sm the used precursors: 1, dtc3Sm bipy; 2, dtc3Sm phen; 3, dtc3Sm (without water ambient); and 4, (with water ambient). (Reprinted with permission from [65].) dtc3Sm (with water ambient). (Reprinted with permission from [65].)

3.5. Other Thin Thin Film Film Deposition Deposition Techniques 3.5. Other Techniques In molecular beam beam epitaxy epitaxy (MBE) (MBE) and and electrodeposition have been been used used for In addition, addition, molecular electrodeposition have for the the deposition of SmS. In the case of MBE, both samarium and sulfur were vaporized on a heated substrate deposition of SmS. In the case of MBE, both samarium and sulfur were vaporized on a heated at 450 ◦ C. at Either SmS or Sm thin films could be obtained by properly changing the sulfur flux [66]. 2 S3 or substrate 450 °C. Either SmS Sm 2S3 thin films could be obtained by properly changing the sulfur Electrodeposition was used to obtain SmS thin films on ITO coated substrates through an flux [66]. aqueous solution whichwas consisted of obtain SmCl3 ·SmS 6H2 Othin and films Na2 S2on O3 ·ITO 5H2 O. The final phase of SmS was Electrodeposition used to coated substrates through an semiconducting, but there was a fraction of Sm S in the films, as indicated by XRD measurements. 2 3 aqueous solution which consisted of SmCl3·6H2O and Na2S2O3·5H2O. The final phase of SmS was ◦ for 30 min, leading to crystallites with both elliptic and circular Annealing took place at 200 semiconducting, but there wasCa fraction of Sm2S3 in the films, as indicated by XRD measurements. shape [67]. took place at 200 °C for 30 min, leading to crystallites with both elliptic and circular shape Annealing

[67]. 3.6. Bulk—Single Crystals 3.6. Bulk—Single Crystals While metallic SmS thin films do not switch back to the semiconducting state upon releasing the pressure, the switching is reversible in single crystals. In the literature, several investigations have While metallic SmS thin films do not switch back to the semiconducting state upon releasing the been published on the growth of single crystal SmS following mainly the same procedure [24,68–71], as pressure, the switching is reversible in single crystals. In the literature, several investigations have proposed by Matsubayashi et al. Initially, they prepared the starting materials (Sm chips and powdered been published on the growth of single crystal SmS following mainly the same procedure [24,68–71], S) placing them in an evacuated quartz ampoule at 600 ◦ C for several days. Afterwards, they used as proposed by Matsubayashi et al. Initially, they prepared the starting materials (Sm chips and an electron beam welding system in which the starting materials (in a tungsten crucible) were fused. powdered S) placing them in an evacuated quartz ampoule at 600 °C for several days. Afterwards, Subsequently, this step was followed by a vertical Bridgman growth in a high frequency induction they used an electron beam welding system in which the starting materials (in a tungsten crucible) oven [68]. The starting material for the Bridgman growth was placed in a tungsten crucible and heated were fused. Subsequently, this step was followed by a vertical Bridgman growth in a high frequency up to 2180 ◦ C (heating rate: 55 ◦ C/h) for 21 h. After that, the sample was cooled down at a rate of induction oven [68]. The starting material for the Bridgman growth was placed in a tungsten crucible 10 ◦ C/h to 1900 ◦ C, obtaining the SmS crystal in the semiconducting phase. and heated up to 2180 °C (heating rate: 55 °C/h) for 21 h. After that, the sample was cooled down at a rate of 10 °C/h to 1900 °C, obtaining the SmS crystal in the semiconducting phase. 4. Applications In this section, we present several applications based on the switching and thermoelectric behavior of SmS. Apart from highlighting the opportunities, we also discuss existing practical difficulties for the final applications. The main reason why the number of real-world applications of

Materials 2017, 10, 953

15 of 21

4. Applications In this section, we present several applications based on the switching and thermoelectric behavior of SmS. Apart from highlighting the opportunities, we also discuss existing practical difficulties for the final applications. The main reason why the number of real-world applications of the switching properties of SmS is still limited, lies is the necessity for high quality single phase SmS films with the correct stoichiometry and Sm valence state. Form the discussion on the different thin film deposition techniques above, it is clear that SmS cannot be straightforwardly obtained, and more research is required in the field. A quite mature application is that of a strain gauge, where SmS could be used, even if in this case there is no direct exploitation of the switching properties of the material. The change in electrical resistivity upon application of stress was monitored by using metallic thin films made by sputtering, as they have shown a strong strain-resistivity relationship. The existence of impurities can influence the pressure-induced resistivity drop of SmS, even in the metallic state [24]. Unfortunately, SmS is rather fragile under stress, and stress overloading can compromise the lifetime of the gauge (with piezoresistive gauge −2.0 GPa−1 ) [24]. A typical structure of a semiconductor strain gauge based on SmS is seen in Figure 12a. The substrate of this device is a lacquer (1), where on top of it a dumbbell shaped constantan carrier (2) is deposited. After that, a SiO (Silicon Monoxide) layer (3) is deposited onto surface 2, as a dielectric layer, with SmS (4) to be the next part of this device. Finally, a nickel layer (5) plays the role of contact [72]. One of the most intriguing applications using the switching behavior of SmS is in the application area of optical storage. If small spots on a SmS surface can be switched in a controlled way, this could be used to store and read information. It was shown that for suitable semiconducting spots (placed on a metallic surface/substrate) with a diameter on the order of 0.5 µm and by using laser pulse energies of 0.5 nJ/(µm)2 , a potential data density of about 50 MB/cm2 can be achieved. The metallic bulk phase could be returned back to the semiconducting phase by local annealing with a pulsed laser. Nevertheless, the shock wave of the laser beam can reconvert the material back to the metallic phase, resulting in a decrease of the density of the optical data storage by a factor of 10 with a simultaneous increase of the required writing energy and the area between the spots. A similar study has been performed by Petrov et al. on holographic data storage, by using nano- and pico-second laser pulses [73]. Unfortunately, technical difficulties appeared. When an intense laser pulse was used, a written spot could be deleted. There is a category of optical storage materials, named phase change materials, which have been studied in previous years, providing important results. Specifically, using these materials, around 2 Gbit/cm2 data storage density can be achieved in optical storage media (Blue Ray disc), a capacity which is still magnitudes higher than what can be obtained in SmS. As a result, for improving the specifications of SmS-based memories, both materials and device improvements should take place [48,74–76]. Another possible application of SmS is as thermoelectric generator. This application is possible thanks to an emerging voltage across a heated SmS bulk material, without the presence of external temperature gradients. This can be observed in the temperature interval from 400 to 500 K. In order to maximize the voltage, a maximum concentration gradient of Sm ions is necessary. A basic structure which has been manufactured, consisted of a Al2 O3 substrate (1), a metallic contact made from nickel (2), on which an overstoichiometric semiconducting Sm1.1 S thin film (3) was subsequently deposited, in the next step an additional semiconducting SmS thin film (4) and finally a metallic contact of Ni (5). When the temperature had reached 428 K an electric voltage equal to 1.1 V appeared, but when the temperature decreased to 360 K, the voltage disappeared (Figure 12b) [77]. In a very similar investigation (on bulk SmS), the specific voltage generation due to the thermoelectric effect could be seen between 100 and 1800 K [78]. Actually, the screening of the Coulomb potential of the additional Sm ions by the electrons of the conduction band plays an important role in the occurrence of the thermoelectric effect, as at higher temperatures there is an activation of electrons coming from the

Materials 2017, 10, 953

16 of 21

4f Sm ion level to the conduction band [78]. Recently, it was described that γ-SmGdEuS4 shows a p-type Seebeck coefficient and high electrical resistivity, providing opportunities for thermoelectric Materials 2017, 10, 16 of 21 applications in953 this class of materials [79].

Figure 12. (a) (a) A A detailed detailed form form of of aa SmS-based SmS-based strain strain gauge gauge sensor sensor (see (see text text for for the the different different parts); parts); (b) Structure of a thermoelectric generator and behavior of voltages and temperature as a function of time; (c) Voltage (at 1 µA μA constant constant current) of SmSe SmSe for three three different different thin thin film film thicknesses. thicknesses. Results were carried out using an indenter (top axis; load in grams) and finite elements calculations (bottom axis); (d) (d) Schematic Schematicillustration illustrationofofthe theproposed proposed PET device. (Reprinted with permission PET device. (Reprinted with permission fromfrom [72],[72], [77] [77] respectively.) and and [80],[80], respectively.)

The The next next step step is is the the manufacturing manufacturing of of ultrathin ultrathin films, films, which which still still demonstrate demonstrate switching switching behavior behavior on would open the way to several applications, with non-volatile electrical storage on the thenanoscale. nanoscale.This This would open the way to several applications, with non-volatile electrical memories as the most important one. This has been confirmed by very recent results storage memories as the most important one. This has been confirmed by very recent resultson on the the piezoresistive behavior of SmSe [80], which could be extended to SmS, promoting a new low-energy piezoresistive behavior of SmSe [80], which could be extended to SmS, promoting a new low-energy memory components. There memory device, device, which which is is based based on on piezoelectric piezoelectric (PE) (PE) and and piezoresistive piezoresistive (PR) (PR) components. There are are three conductive parts (gate, common, sense) in this device. The PE material, which is placed three conductive parts (gate, common, sense) in this device. The PE material, which is placed between between the and common, common, expands expands upon application of voltage, triggering triggering aa semiconductor–metal semiconductor–metal the gate gate and upon application of aa voltage, transition in the PR layer that is placed between the common and sense. All of these components are transition in the PR layer that is placed between the common and sense. All of these components mounted in a yoke, which has a high Young modulus (Figure 12d). In Figure 12c, we see the voltage are mounted in a yoke, which has a high Young modulus (Figure 12d). In Figure 12c, we see the (proportional to resistance) of threeofSmSe films a function of applied pressure. The top voltage (proportional to resistance) threethin SmSe thinas films as a function of applied pressure. Theaxis top represents the load of an indenter (in grams), while the bottom axis shows the corresponding pressure axis represents the load of an indenter (in grams), while the bottom axis shows the corresponding from calculations (finite element calculations; in GPa). The The change in voltage is isclearly pressure from calculations (finite element calculations; in GPa). change in voltage clearlyseen, seen, by by applying pressure [80]. In this kind of application, SmS can provide a hysteretic behavior, which applying pressure [80]. In this kind of application, SmS can provide a hysteretic behavior, which does does not not take take place place in in the the case case of of SmSe. SmSe. This This would would decrease decrease at at the the same same time time the the required required input input voltage, voltage, which is responsible responsiblefor forthe theexpansion expansionofofthe thePE PEpart part the device and a result compression of which is ofof the device and as as a result forfor compression of the the PR material, which will change between a low and a high resistance state (metal–insulator PR material, which will change between a low and a high resistance state (metal–insulator transition) transition) translating these changes to 0 and 1. In this case, it is a challenge to modify the composition translating these changes to 0 and 1. In this case, it is a challenge to modify the composition in such a in such a way that the transition back to the semiconducting state can be achieved without annealing way that the transition back to the semiconducting state can be achieved without annealing in vacuum. in vacuum. Alloying with other elements, such as EuS and GdS can be used to induce a repeatable, reversible switching between the states. This idea can promote an appropriate hysteretic resistivity loop as a function of pressure, provided clamping effects of substrate and other components are properly taken into account.

Materials 2017, 10, 953

17 of 21

Alloying with other elements, such as EuS and GdS can be used to induce a repeatable, reversible switching between the states. This idea can promote an appropriate hysteretic resistivity loop as a function of pressure, provided clamping effects of substrate and other components are properly taken into account. 5. Conclusions In this review, material properties and possible applications of samarium sulfide were reviewed. SmS is a switchable material, which shows a first order transition (semiconductor–metal) at about 0.65 GPa, characterized by a volume collapse without a change in crystallographic phase, and an additional magnetic transition at about 1.9 GPa. SmS presents many intriguing phenomena such as Kondo-like behavior and the existence of a pseudo-gap before the final closing of its band gap, when reaching the metallic phase. Different experimental and computational techniques were used to study its switchable properties, giving at the same time additional information about practical applications. A crucial issue is the presence of the suitable oxidation state of Sm, which is responsible for the switching behavior. We have to note that SmS bulk crystals properties have been more thoroughly studied in comparison with SmS thin films, although for thin films the influence of stress, induced by the substrate, can strongly alter the switching behavior. Obtaining the correct stoichiometry in SmS can be an experimental challenge, although several techniques are described in literature, such as pulsed laser deposition, electron beam evaporation and sputtering. By alloying, for instance by europium or gadolinium, the resistivity and the switching behavior can be altered. While the switching in SmS was initially dominantly studied from a fundamental point of view, the first application interests were in strain sensing and optical memory devices, as, in the latter case, laser irradiation can also induce the switching. Recently, non-volatile memory devices based on combined piezoelectric and piezoresistive materials have been proposed. In summary, SmS can be considered as an intriguing switchable material, as it shows a pressure-induced semiconductor–metal transition at moderate pressure, at room temperature. Nevertheless, the need for high-quality SmS thin films in view of applications simultaneously boosts the study of fundamental properties of SmS thin films. This will allow the usage of this material in high-quality sensing and memory devices, as well as very presumably in optical switching devices, due to the laser-induced transition of SmS. These applications will promote a further study not only of SmS thin films, but also of other pressure-induced insulator–metal compounds. Acknowledgments: This project has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 688282. Author Contributions: Andreas Sousanis made the films and experiments needed for Figure 3c and wrote the first draft of the paper. Both Philippe F. Smet and Dirk Poelman extensively restructured, rewrote and proofread the paper. Conflicts of Interest: The authors declare no conflict of interest.

References 1.

2. 3. 4. 5.

Burr, G.W.; Breitwisch, M.J.; Franceschini, M.; Garetto, D.; Gopalakrishnan, K.; Jackson, B.; Kurdi, B.; Lam, C.; Lastras, L.A.; Padilla, A.; et al. Phase change memory technology. J. Vac. Sci. Technol. B 2010, 28, 223–262. [CrossRef] Lencer, D.; Salinga, M.; Wuttig, M. Design rules for phase-change materials in data storage applications. Adv. Mater. 2011, 23, 2030–2058. [CrossRef] [PubMed] Lencer, D.; Salinga, M.; Grabowski, B.; Hickel, T.; Neugebauer, J.; Wuttig, M. A map for phase-change materials. Nat. Mater. 2008, 7, 972–977. [CrossRef] [PubMed] Chernova, N.A.; Roppolo, M.; Dillon, A.C.; Whittingham, M.S. Layered vanadium and molybdenum oxides: Batteries and electrochromics. J. Mater. Chem. 2009, 19, 2526–2552. [CrossRef] Fortier, J.P.; Baloukas, B.; Zabeida, O.; Klemberg-Sapieha, J.E.; Martinu, L. Thermochromic VO2 thin films deposited by HiPIMS. Sol. Energy Mater. Sol. Cells 2014, 125, 291–296. [CrossRef]

Materials 2017, 10, 953

6. 7. 8.

9.

10. 11. 12. 13. 14. 15.

16. 17.

18.

19. 20. 21. 22.

23. 24. 25. 26.

27. 28.

18 of 21

Li, S.Y.; Niklasson, G.A.; Granqvist, C.G. Thermochromic fenestration with VO2 -based materials: Three challenges and how they can be met. Thin Solid Films 2012, 520, 3823–3828. [CrossRef] Maaza, M.; Nemraoui, O.; Sella, C.; Beye, A.C.; Baruch-Barak, B. Thermal induced tunability of surface plasmon resonance in Au-VO2 nano-photonics. Opt. Commun. 2005, 254, 188–195. [CrossRef] Jarrige, I.; Yamaoka, H.; Rueff, J.P.; Lin, J.F.; Taguchi, M.; Hiraoka, N.; Ishii, H.; Tsuei, K.D.; Imura, K.; Matsumura, T.; et al. Unified understanding of the valence transition in the rare-earth monochalcogenides under pressure. Phys. Rev. B 2013, 87, 115107. [CrossRef] Sousanis, A.; Smet, P.F.; Detavernier, C.; Poelman, D. Stability of switchable SmS for piezoresistive applications. In Proceedings of the IEEE Nanotechnol Materials and Devices Conference (NMDC), Toulouse, France, 9–12 October 2016. Kitagawa, R.; Takebe, H.; Morinaga, K. Photoinduced phase transition of metallic SmS thin films by a femtosecond laser. Appl. Phys. Lett. 2003, 82, 3641. [CrossRef] Suryanarayanan, R.; Smirnov, I.A.; Brun, G.; Shul’Man, S.G. Optical absorption of semiconducting and metallic SmS nd Sm1-x Lnx S films. J. Phys. Colloq. 1976, 37, 271–274. [CrossRef] Hickey, C.F. Optical, chemical, and structural properties of thin films of samarium-sulfide and zinc-sulfide. Ph.D. Thesis, University of Arizona, Tucson, AZ, USA, 1987. Imura, K.; Matsubayashi, K.; Suzuki, H.S.; Kabeya, N.; Deguchi, K.; Sato, N.K. Pressure-temperature phase diagram of golden SmS. J. Phys. Soc. Jpn. 2009, 78, 104602. [CrossRef] Jayaraman, A.; Bucher, E.; Dernier, P.D.; Longinotti, L.D. Temperature-induced explosive first-order electronic phase-transition in Gd-doped Sms. Phys. Rev. Lett. 1973, 31, 700–703. [CrossRef] Rogers, E.; Smet, P.F.; Dorenbos, P.; Poelman, D.; van der Kolk, E. The thermally induced metal-semiconducting phase transition of samarium monosulfide (SmS) thin films. J. Phys. Condens. Matter 2010, 22, 015005. [CrossRef] [PubMed] Maple, M.B.; Wohllebe, D. Nonmagnetic 4f shell in high pressure phase of Sms. Phys. Rev. Lett. 1971, 27, 511–515. [CrossRef] Ariponnammal, S.; Chandrasekaran, S. Theoretical study on the electrical properties of some semiconducting rare earth chalcogenides Sm1−x Eux S and Sm1−x Ybx S under pressure. J. Nano- Electron. Phys. 2011, 3, 529–535. Menushenkov, A.P.; Chernikov, R.V.; Sidorov, V.V.; Klementiev, K.V.; Alekseev, P.A.; Rybina, A.V. Relationship between the local electronic and local crystal structures of intermediate-valence Sm1−x Yx S. JETP Lett. 2006, 84, 119–123. [CrossRef] Pergament, A. Metal-insulator transition: The Mott criterion and coherence length. J. Phys. Condens. Matter 2003, 15, 3217–3223. [CrossRef] Pergament, A.; Kazakova, E.; Morak, A. The photoinduced Mott transition from metal to insulator: The problem of critical concentration. J. Phys. Condens. Matter 2005, 17, 1151–1156. [CrossRef] Kampfer, T. Die dynamik des laserinduzierten metall-halbleiter-phasenubergangs von samariumsulfid (SmS). Ph.D. Thesis, University of Jena, Jena, Germany, 2009. Barla, A.; Sanchez, J.P.; Haga, Y.; Lapertot, G.; Doyle, B.P.; Leupold, O.; Ruffer, R.; Abd-Elmeguid, M.M.; Lengsdorf, R.; Flouquet, J. Pressure-induced magnetic order in golden SmS. Phys. Rev. Lett. 2004, 92, 066401. [CrossRef] [PubMed] Kim, D.J.; Xia, J.; Fisk, Z. Topological surface state in the Kondo insulator samarium hexaboride. Nat. Mater. 2014, 13, 466–470. [CrossRef] [PubMed] Imura, K.; Matsubayashi, K.; Suzuki, H.S.; Deguchi, K.; Sato, N.K. Thermodynamic and transport properties of SmS under high pressure. Physica B 2009, 404, 3028–3031. [CrossRef] Jayaraman, A.; Maines, R.G. Study of the valence transition in Eu-, Yb-, and Ca-substituted SmS under high pressure and some comments on other substitutions. Phys. Rev. B 1979, 19, 4154–4161. [CrossRef] Deen, P.P.; Braithwaite, D.; Kernavanois, N.; Paolasini, L.; Raymond, S.; Barla, A.; Lapertot, G.; Sanchez, J.P. Structural and electronic transitions in the low-temperature, high-pressure phase of SmS. Phys. Rev. B 2005, 71, 245118. [CrossRef] Wang, L.; Marin, C.M.; Mei, W.-N.; Li Cheung, C. Electronic structures of lanthanum, samarium, and gadolinium sulfides. AIMS Mater. Sci. 2015, 2, 97–105. [CrossRef] Wachter, P. Handbook of the Physical and Chemistry of Rare Earths. In Handbook of the Physical and Chemistry of Rare Earths; Elsevier: Amsterdam, The Netherlands, 1993; Volume 19.

Materials 2017, 10, 953

29.

30. 31.

32.

33. 34. 35. 36. 37. 38.

39. 40.

41. 42. 43.

44. 45.

46. 47. 48. 49. 50.

19 of 21

Rogers, E.; Dorenbos, P.; Haas, J.T.M.D.; Kolk, E.V.d. Experimental study of the 4fn →4fn and 4fn →4fn−1 5d1 transitions of the lanthanide diiodides LnI2 (Ln = Nd, Sm, Eu, Dy, Tm, Yb). J. Phys. Condens. Matter 2012, 24, 275502. [CrossRef] [PubMed] Batlogg, B.; Kaldis, E.; Schlegel, A.; Wachter, P. Electronic structure of Sm monochalcogenides. Phys. Rev. B 1976, 14, 5503. [CrossRef] Van der Kolk, E.; Dorenbos, P. Systematic and material independent variation of electrical, optical, and chemical properties of Ln-materials over the Ln-series (Ln = La, Ce, Pr, ..., Lu). Chem. Mater. 2006, 18, 3458–3462. [CrossRef] Imura, K.; Hajiri, T.; Matsunami, M.; Kimura, S.; Kaneko, M.; Ito, T.; Nishi, Y.; Sato, N.K.; Suzuki, H.S. Angle-resolved photoemission spectroscopy on mixed-valent Sm1−x Yx S. J. Korean Phys. Soc. 2013, 62, 2028–2031. [CrossRef] Binnemans, K.; Van Deun, R.; Görller-Walrand, C.; Adam, J.L. Spectroscopic properties of trivalent lanthanide ions in fluorophosphate glasses. J. Non-Cryst. Solids 1998, 238, 11–29. [CrossRef] Kimura, S.I.; Mizuno, T.; Matsubayashi, K.; Imura, K.; Suzuki, H.S.; Sato, N.K. Infrared study on the electronic structure of SmS in the black phase. Physica B 2008, 403, 805–807. [CrossRef] Van Deun, R.; Binnemans, K.; Gorller-Walrand, C.; Adam, J.L. Spectroscopic properties of trivalent samarium ions in glasses. Rare-Earth-Doped Mater. Devices III 1999, 3622, 175–181. [CrossRef] Matsubayashi, K.; Imura, K.; Suzuki, H.S.; Ban, S.; Chen, G.F.; Deguchi, K.; Sato, N.K. Magnetic properties of golden SmS. J. Magn. Magn. Mater. 2007, 310, 408–410. [CrossRef] Kikoin, K.A. Nature of the “golden” phase of samarium monosulfide. Zh Eksp Teor Fiz+ 1983, 85, 1000–1016. Shlimak, I.; Kaveh, M.; Ussyshkin, R.; Ginodman, V.; Resnick, L.; Gantmakher, V.F. Quantitative analysis of delocalization in the vicinity of the metal-insulator transition in doped semiconductors. J. Phys.Condens. Matter 1997, 9, 9873–9880. [CrossRef] Kirk, J.L.; Vedam, K.; Narayanamurti, V.; Jayaraman, A.E.B. Direct optical observation of the semiconductor to metal transition in SmS under pressure. Phys. Rev. B 1972, 6, 3023. [CrossRef] Mizuno, T.; Iizuka, T.; Kimura, S.; Matsubayashi, K.; Imura, K.; Suzuki, H.S.; Sato, N.K. Excitonic instability in the transition from the black phase to the golden phase of SmS under pressure investigated by infrared spectroscopy. J. Phys. Soc. Jpn. 2008, 77, 113704. [CrossRef] Hickey, C.F.; Gibson, U.J. Optical response of switching SmS in thin films prepared by reactive evaporation. J. Appl. Phys. 1987, 62, 3912. [CrossRef] Imura, K.; Saito, M.; Kaneko, M.; Ito, T.; Hajiri, T.; Matsunami, M.; Kimura, S.; Deguchi, K.; Suzuki, H.S.; Sato, N.K. Origin of the black-golden transition in Sm1−x Yx S. J. Phys. Conf. Ser. 2015, 592. [CrossRef] Sharenkova, N.V.; Kaminskii, V.V.; Golubkov, A.V.; Romanova, M.V.; Stepanov, N.N. Mechanism of stabilization of the Sm1−x Gdx S metallic modification at the pressure-induced semiconductor-metal phase transition. Phys. Solid State 2009, 51, 1700–1702. [CrossRef] Chaboy, J.; Garcia, J.; Marcelli, A. Correlation between mixed valence behavior of cerium and the magnetic and superconducting phenomena of CeFe2 and CeRu2 . J. Magn. Magn. Mater. 1992, 104, 661–662. [CrossRef] Alekseev, P.A.; Chernikov, R.V.; Golubkov, A.V.; Klementiev, K.V.; Menushenkov, A.P.; Nemkovsky, K.S. XAFS spectroscopy of the mixed valent Sm1−x Yx S. Nucl. Instrum. Methods Phys. Res. Sect. A Accel. Spectrom. Detect. Assoc. Equip. 2005, 543, 205–207. [CrossRef] Mori, Y.; Tanemura, S. Chemical analysis of semiconducting and metallic SmS thin films by X-ray photoelectron spectroscopy. Appl. Surf. Sci. 2007, 253, 3856–3859. [CrossRef] Mori, Y.; Tanemura, S.; Koide, S.; Senzaki, Y.; Jin, P.; Kaneko, K.; Terai, A.; Nabotova-Gabin, N. Thermo-coloration of SmS thin flims by XPS in situ observation. Appl. Surf. Sci. 2003, 212, 38–42. [CrossRef] De Tomasi, F.; Perrone, M.R.; Protopapa, M.L.; Leo, G. Laser irradiation effects on the resistance of SmS films. Thin Solid Films 2002, 413, 171–176. [CrossRef] Miodushevsky, P.; Protopapa, M.L.; De Tomasi, F.; Perrone, M.R.; Tundo, S.; Vasanelli, L. Fine trimming of SmS film resistance by XeCl laser ablation. Thin Solid Films 2000, 359, 251–254. [CrossRef] Imura, K.; Kanematsu, S.; Matsubayashi, K.; Suzuki, H.S.; Deguchi, K.; Sato, N.K. Discontinuous transition from a real bound state to virtual bound state in a mixed-valence state of SmS. J. Phys. Soc. Jpn. 2011, 80, 113704. [CrossRef]

Materials 2017, 10, 953

51.

52.

53. 54. 55.

56. 57.

58. 59. 60. 61. 62. 63.

64. 65.

66. 67. 68. 69. 70. 71. 72. 73.

20 of 21

Kang, C.J.; Choi, H.C.; Kim, K.; Min, B.I. Topological properties and the dynamical crossover from mixed-valence to Kondo lattice behavior in the golden phase of SmS. Phys. Rev. Lett. 2015, 114, 166404. [CrossRef] [PubMed] Barla, A.; Sanchez, J.P.; Derr, J.; Salce, B.; Lapertot, G.; Flouquet, J.; Doyle, B.P.; Leupold, O.; Ruffer, R.; Abd-Elmeguid, M.M.; et al. Valence and magnetic instabilities in Sm compounds at high pressures. J. Phys. Condens. Matter 2005, 17, 837–848. [CrossRef] Zhao, Y.C.; Porsch, F.; Holzapfel, W.B. Intermediate 4f bonding structure for samarium under pressure. Phys. Rev. B 1994, 50, 6603–6608. [CrossRef] Haga, Y.; Derr, J.; Barla, A.; Salce, B.; Lapertot, G.; Sheikin, I.; Matsubayashi, K.; Sato, N.K.; Flouquet, J. Pressure-induced magnetic phase transition in gold-phase SmS. Phys. Rev. B 2004, 70, 220406. [CrossRef] Takahashi, H.; Okazaki, R.; Taniguchi, H.; Terasaki, I.; Saito, M.; Imura, K.; Deguchi, K.; Sato, N.K.; Suzuki, H.S. Electrical oscillation in SmS induced by a constant external voltage. Phys. Rev. B 2014, 89, 195103. [CrossRef] Witczak-Krempa, W.; Chen, G.; Kim, Y.B.; Balents, L. Correlated Quantum Phenomena in the Strong Spin-Orbit Regime. Ann. Rev. Condens. Matter Phys. 2014, 5, 57–82. [CrossRef] Shukla, N.; Parihar, A.; Freeman, E.; Paik, H.; Stone, G.; Narayanan, V.; Wen, H.; Cai, Z.; Gopalan, V.; Engel-Herbert, R.; et al. Synchronized charge oscillations in correlated electron systems. Sci. Rep. 2014, 4, 4964. [CrossRef] Li, Z.; Li, J.; Blaha, P.; Kioussis, N. Predicted topological phase transition in the SmS Kondo insulator under pressure. Phys. Rev. B 2014, 89, 121117. [CrossRef] Wachter, P.; Jung, A.; Steiner, P. Pressure-driven metal-insulator transition in La-doped SmS—Excitonic Condensation. Phys. Rev. B 1995, 51, 5542–5545. [CrossRef] Stern, A.; Dzero, M.; Galitski, V.M.; Fisk, Z.; Xia, J. Surface-dominated conduction up to 240 K in the Kondo insulator SmB6 under strain. Nat. Mater. 2017. [CrossRef] [PubMed] Rogers, E. Engineering the Electronic Structure of Lanthanide Based Materials. Ph.D. Thesis, TUDelft, Delft, The Netherland, 2012. Jin, P.; Tazawa, M.; Huang, J.F.; Tanemura, S. Growth of samarium monosulfide thin films by co-sputtering deposition. J. Cryst. Growth 1998, 191, 285–289. [CrossRef] Tanemura, S.; Koide, S.; Senzaki, Y.; Miao, L.; Hirai, H.; Mori, Y.; Jin, P.; Kaneko, K.; Terai, A.; Nabatova-Gabain, N. Fabrication and characterization of metal and semiconductor SmS thin films by rf/dc dual magnetron sputtering. Appl. Surf. Sci. 2003, 212, 279–286. [CrossRef] Zenkevich, A.V.; Parfenov, O.E.; Storchak, V.G.; Teterin, P.E.; Lebedinskii, Y.Y. Highly oriented metallic SmS films on Si(100) grown by pulsed laser deposition. Thin Solid Films 2011, 519, 6323–6325. [CrossRef] Volodin, N.M.; Zavyalova, L.V.; Kirillov, A.I.; Svechnikov, S.V.; Prokopenko, I.V.; Khanova, A.V. Investigation of growth conditions, crystal structure and surface morphology of SmS films fabricated by MOCVD technique. Semicond. Phys. Quantum Electron. Optoelectron. 1999, 2, 78–83. Suryanarayanan, R.; Brun, G. A compact multi-source multi-substrate evaporator for thin film studies of rare earth sulphides. Thin Solid Films 1976, 35, 263–271. [CrossRef] Huang, J.F.; Ma, X.B.; Cao, L.Y.; Wu, J.P.; He, H.Y. Electrodeposition of SmS optical thin films on ITO glass substrate. Mater. Lett. 2007, 61, 3920–3922. [CrossRef] Matsubayashi, K.; Suzuki, H.S.; Imura, K.; Nishioka, T.; Sato, N.K. Single crystal growth of “pure” SmS. Physica B 2005, 359, 151–153. [CrossRef] Kaminskii, V.V.; Solov’ev, S.M. Thermal effects in homogeneously heated samarium sulfide single crystals. Tech. Phys. Lett. 2005, 31, 603–604. [CrossRef] Imura, K.; Matsubayashi, K.; Suzuki, H.S.; Nishioka, T.; Mori, N.; Sato, N.K. Thermal expansion study on high-pressure phases of SmS. Physica B 2006, 378–380, 728–729. [CrossRef] Imura, K.; Matsubayashi, K.; Suzuki, H.S.; Deguchi, K.; Sato, N.K. Transport properties of golden SmS. Physica B 2008, 403, 895–897. [CrossRef] Volodin, N.M.; Mishin, Y.N.; Kaminskii, V.V.; Zakharov, Y.V. Samarium monosulfide-based semiconductor strain gages for spacecraft-strain transformation. Sol. Syst. Res. 2013, 47, 601–604. [CrossRef] Petrov, M.P.; Grachev, A.I.; Kukharskii, A.A.; Smirnov, I.A.; Shulman, S.G. Holographic storage in Sms thin films. Opt. Commun. 1977, 22, 293–296. [CrossRef]

Materials 2017, 10, 953

74. 75. 76. 77. 78. 79.

80.

21 of 21

Welnic, W.; Pamungkas, A.; Detemple, R.; Steimer, C.; Blugel, S.; Wuttig, M. Unravelling the interplay of local structure and physical properties in phase-change materials. Nat. Mater. 2006, 5, 56–62. [CrossRef] Wuttig, M.; Raoux, S. The Science and Technology of Phase Change Materials. Z. Anorg. Allg. Chem. 2012, 638, 2455–2465. [CrossRef] Meena, J.S.; Sze, S.M.; Chand, U.; Tseng, T.Y. Overview of emerging nonvolatile memory technologies. Nanoscale Res. Lett. 2014, 9, 526. [CrossRef] [PubMed] Kaminskii, V.V.; Kazanin, M.M. Thermovoltaic effect in thin-film samarium-sulfide-based structures. Tech. Phys. Lett. 2008, 34, 361–362. [CrossRef] Kaminskii, V.V.; Didik, V.A.; Kazanin, M.M.; Romanova, M.V.; Solov’ev, S.M. Thermovoltaic effect in polycrystalline samarium sulfide. Tech. Phys. Lett. 2009, 35, 981–984. [CrossRef] Ohta, M.; Kuzuya, T.; Sasaki, H.; Kawasaki, T.; Hirai, S. Synthesis of multinary rare-earth sulfides PrGdS3 , NdGdS3 , and SmEuGdS4 , and investigation of their thermoelectric properties. J. Alloys Compd. 2009, 484, 268–272. [CrossRef] Copel, M.; Kuroda, M.A.; Gordon, M.S.; Liu, X.H.; Mahajan, S.S.; Martyna, G.J.; Moumen, N.; Armstrong, C.; Rossnagel, S.M.; Shaw, T.M.; et al. Giant piezoresistive On/Off ratios in rare-earth chalcogenide thin films enabling nanomechanical switching. Nano Lett. 2013, 13, 4650–4653. [CrossRef] [PubMed] © 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Cloud-Based Applications for Organizing and Reviewing Plastic Surgery Content.

Nanoemulsions: formation, properties and applications.

Properties and applications of pyrethroids.

Cholesterol oxidases: properties and applications.

Reviewing Einstein.

Fluorinated sulfoximines: syntheses, properties and applications.

Cellulose nanocrystals: synthesis, functional properties, and applications.

Design, properties, and applications of neutral ionophores.

Thin metal nanostructures: synthesis, properties and applications.

Diamond nanowires: fabrication, structure, properties, and applications.

Multispectral phloem-mobile probes: properties and applications.

Halogenated boron-dipyrromethenes: synthesis, properties and applications.

Psychometric Properties of the "Sport Motivation Scale (SMS)" Adapted to Physical Education.

Charge for peer reviewing.

The ethics of reviewing.

Reviewing an appointment system.

Reviewing physicians' practices.

Reviewing multiple sclerosis.

Appraising journals' reviewing procedures.

Reviewing the reviewers.

Existing MRC reviewing procedure.

Reviewing the Mental Health Act.

Cross-coupling reactions using samarium(II) iodide.

Reviewing post-publication peer review.